Effect of pathologic revision and Ki67 and ERG immunohistochemistry on predicting radical prostatectomy outcome in men initially on active surveillance.

OBJECTIVE: To investigate if pathologic biopsy reevaluation and implementation of immunohistochemical biomarkers could improve prediction of radical prostatectomy outcome in men initially on active surveillance.

METHODS: Biopsy specimens from diagnosis until switching to radical prostatectomy in men initially on active surveillance in the Dutch part of the Prostate cancer Research International Active Surveillance (PRIAS) study were collected and revised by a single pathologist. Original and revised biopsy Gleason score were compared and correlated with radical prostatectomy Gleason score. Biopsy specimens were immunohistochemically stained for Ki67 and ERG. Predictive ability of clinical characteristics and biomarkers on Gleason ≥7 or ≥pT3 on radical prostatectomy was tested using logistic regression and ROC curve analysis.

RESULTS: A total of 150 biopsies in 95 men were revised. In 13% of diagnostic or second-to-last biopsies and 20% of the last biopsies on active surveillance revision of Gleason score resulted in change of recommendation (ie, active treatment or active surveillance). Concordance with Gleason score on radical prostatectomy was however similar for both the revised and original Gleason on biopsy. Ki67 and ERG were not statistically significant predictors of Gleason ≥7 or ≥pT3 on radical prostatectomy.

CONCLUSIONS: Although interobserver differences in pathology reporting on biopsy could result in a change of management strategy in approximately 13-20% of men on active surveillance, both pathological revision and tested biomarkers (Ki67 and ERG) did not improve prediction of outcome on radical prostatectomy. Undersampling of most aggressive tumor remains the main focus in order to increase accurate grading at time of treatment decision making.